Getting Help for your Biostatistics Questions &

Database Basics

Dan Eastwood, MS, Biostatistician Medical College of Wisconsin, Division of Biostatistics

Friday, October 5, 2012

12:00-1:00 pm

The Medical College of Wisconsin is accredited by the Accreditation Council for Continuing Medical Education to provide continuing

medical education for physicians.

The Medical College of Wisconsin designates this Live activity for a maximum of 1.0 AMA PRA Category 1 Credit(s) ™.

Physicians should claim only the credit commensurate with the extent of their participation in the activity.

Hours of Participation for Allied Health Professionals

The Medical College of Wisconsin designates this activity for up to 1.0 hours of participation for continuing education for allied health

professionals.

2

Financial Disclosure • In accordance with the ACCME® standard for Commercial

Support Number 6, all in control of content disclosed any relevant financial relationships. The following in control of content had no relevant financial relationships to disclose.

Name: Role in Meeting:

Ruta Brazauskas, PhD Planning Committee

Haley Montsma, BBA Planning Committee

Dan Eastwood, MS Speaker

3

Learning Objectives • Discover the capabilities and resources available within

the Biostatistics Consulting Service

• Transition your research idea into a testable hypothesis • Effectively organize research data

4

Evaluation Forms Your opinion matters!

Help us plan future meetings, by completing and submitting your evaluation forms.

Thank you.

5

Getting Help for your Biostatistics Questions

Dan Eastwood, MS

6

Understanding Statistics

XKCD.com

7

Why Biostatistics Consulting?

• Shared experience • Discuss your study • Consider alternate views • Formulate ideas into hypotheses

8

Why Biostatistics Consulting?

Why might you need help? • “I’ve got this research idea about …” • … • … • … • … • “The reviewers asked me to …”

9

Why Biostatistics Consulting?

Why might you need help? • “I’ve got this research idea about …” • “How large should my sample be?” • … • … • “A significant result! What does it mean?” • “The reviewers asked me to …”

10

Why Biostatistics Consulting?

Why might you need help? • “I’ve got this research idea about …” • “How large should my sample be?” • “I need help organizing my data.” • “How do I perform a Chi-square test?” • “A significant result! What does it mean?” • “The reviewers asked me to …”

11

Questions for the Investigator • What is your hypothesis? • What is the experimental

design? • What type of data are

available? • What is the plan for analysis?

12

Questions for the Investigator What makes a good hypothesis? (1) • A simple sentence, from which the null and

alternate hypothesis should be clear • The alternate hypothesis should be reasonable

(power and effect size) • Difference, equivalence, agreement • No hypothesis- a descriptive study

13

Questions for the Investigator What makes a good hypothesis? (2) • Statistical and clinical significance • Related factors • Preliminary data

14

It Never Hurts to Ask

Independent Samples

Sleep difficulty,

Medication X?

Sleep difficulty,

Medication Y? Total

“No” 48 78 126

“Yes” 136 106 242

Total 184 184 368 Chi-Square test, p=0.0010, Medication X 57.6%, Medication Y 73.9% 15

It Never Hurts to Ask Paired Data

Med Y, No difficulties

Med Y, Sleep Difficulties Total

Med X, No difficulties

34

44

78

Med X, Sleep Difficulties

14

92 106

Total 48 136 184 McNemar’s Test, p=0.0001, paired data odds ratio = 44/14 = 3.14 16

Questions for the Investigator

What data resources are available? • Understand your data • Clinical data and public databases • Data management

• Good data = good research

17

The Answers

What is the plan for analysis? • Best methods for the available data • Best data for the available methods • Potential for other analyses

18

Biostatistics Consulting Service

How to find us: • Schedule a meeting. • Just “Drop-In”. • Special Sessions. • Ask Us.

19

Consulting Services Faculty • John Klein, PhD, Professor & Director of Division of

Biostatistics, Fellow ASA • Survival analysis, cancer statistics, gait analysis

• Aniko Szabo, PhD, Associate Professor & BCS Director

• Cancer statistics, genetics, clinical trials

• Sergey Tarima, PhD, Assistant Professor • Missing data problems, health service research

• Tao Wang, PhD, Associate Professor

• Statistical genetics

• Jessica Pruszynski, PhD, Assistant Professor • Logistic regression

20

Consulting Services Staff • Dan Eastwood, MS, BCS Manager

• Gait analysis, health policy, general biostatistics

• Alexis Visotcky, MS, Biostatistician • VA databases, REDCap

• Qun (Katelyn) Xiang, MS, Biostatistician • Large databases, pediatric data

• Shi (Heather) Zhao, MS, Biostatistician • Haley Montsma, BBA, Administrator

21

What should you bring to a meeting?

• Ideas • Protocol? • Example of your data • Electronic copy of your data? • “The boss”

22

Services Data entry (fee service). Help with: • Design • Analysis • Grant Preparation • Reading Papers • Reports • Graphics

23

• Assistance with Public

Databases • Advice on Methods

Biostatistics Consulting Service • We are now supported by the Medical College’s

Clinical and Translational Science Institute (CTSI) • Biostatistics key function • Monthly Lecture Series (more stats!):

www.mcw.edu/biostatistics/LectureSeries.htm

• DATUM newsletter: www.mcw.edu/biostatistics/datum.htm

24

25

Biostatistics Consulting Service CTSI services available to faculty, staff, and students working on Clinical and Translational Science Research at: • MCW • VA Medical Center • Blood Center • UW-Milwaukee • Marquette • Milwaukee School of Engineering

Free Drop-in Consulting • Medical College of Wisconsin:

Tuesdays and Thursdays Time: 1:00 PM—3:00 PM Building: Health Research Center Room: H2400 Biostatistics

• MCW Cancer Center Wednesdays 10:00 AM—12:00 PM Fridays 1:00 PM—3:00 PM Building: MCW Clinical Cancer Center Room: Clinical Trials Support Room CLCC: 3236 (Enter through C3233)

• Froedtert Pavilion: Mondays & Wednesdays Time: 1:00 PM—3:00 PM Building: Froedtert Pavilion Room: L772A- TRU Offices (Lower Level)

• Clement J. Zablocki VA Medical Center: 1st & 3rd Monday of the month Time: 9:00 AM—11:00 AM Building: 111, 5th Floor B-wing Room: 5423

• Marquette University: Every Tuesday Time: 8:30 AM—10:30 AM Building: School of Nursing—Clark Hall Room: Office of Research and Scholarship: 112D Contact: Jessica Pruszynski, PhD to make an appointment Please note: Priority given to MU Nursing and Dental School personnel

26

Contact

• Haley Montsma • (414) 955-7439 • hmontsma@mcw.edu • consult@mcw.edu

• Dan Eastwood, MS • (414) 955-4855 • eastwood@mcw.edu • consult@mcw.edu

www.mcw.edu/biostatsconsult.htm

27

Database Basics

Dan Eastwood, MS

28

What is a Database? • An organized collection of data

• Accessible in a computer

• Accessible in various ways

• sortable • searchable • indexed

29

What is a Database?

• Well organized data enables good research

• Complex studies require careful organization

• Simple studies benefit from good organization

30

Is a spreadsheet a database?

31

Spreadsheet vs. Database

• Spreadsheets have few or no rules

• Databases have strict rules

• Rules make spreadsheets more like a database

32

Spreadsheet vs. Database

• Use a database program to enforce rules

• Additional capability of databases

• A simple database can be viewed in spreadsheet or “table” form

33

Spreadsheet vs. Database

• Spreadsheets are prone to copy/paste, partial sorting, and other entry errors • Errors may be uncorrectable • Errors may be undetectable

• Changes to databases are generally reversible • Queries display data in different ways • Revert to original • Errors are more easily detected

34

What goes into a database?

• Type of data, formatting • The “bad” list • Factors and variables • Sample units or observations? • Multiple tables and linked tables

35

Statistical Qualities of Data

36

Computational Qualities of Data

37

The List of Bad Things • More than one value in a single cell • Mixed character and numbers data • Merged cells • Color coding • UPPER and lower case text are different • Confused coding or formats • “Prettifying” is generally unhelpful • Identifying information (try to minimize)

38

39

Better now?

40

Variables • Short yet meaningful name

• top row of spreadsheet • longer description or label

• Create a “key” to formatted values

• ex: 1=yes, 2=no 1=treatment, 0=control • usually on a different sheet

41

Factors & Variables • A factor is a complete description of one

contributing element in the analysis • A variable is a representation of a factor, or

part of a factor, as used in the analysis • A factor may be described by several variables

42

Factors & Variables

43

Sample Units & Observations • Depends on Study Design • Usually one row of data per sample unit

• ie: one row per patient • “wide” layout • side-to-side scrolling problems

• Sometimes one row per observation • “long” layout • wasted space with demographics

44

Wide Layout Form

45

Long Layout Form

46

Multiple Tables - Demographics and Clinical Data with Linking index variable

47

Database Programs • Microsoft Access

• Everybody has it • Nobody uses it

48

Database Programs • REDCap

• web based • secure server • survey package (no more Survey Monkey) • 267 institutional partners • 20K+ studies, 30K+ end users • project-redcap.org

• Contact for more information: • Mark Oium, moium@mcw.edu, 805-2051 49

Concluding Remarks • Have a plan for your data • You can “pilot” a database at the same time you

gather pilot data for a study • Good data leads to good research

50

Questions?

51